Heterobimetallic bismuth-transition metal salicylate complexes as molecular precursors for ferroelectric materials. Synthesis and structure

Article abstract of DOI:10.1021/ic0255454

The reactions between triphenylbismuth, salicylic acid, and the metal alkoxides M(OCH₂CH₃)₅ (M = Nb, Ta) or Ti{OCH(CH₃)₂}₄ have been investigated under different reaction conditions and in

Full text of DOI:10.1021/ic0255454

Inorg. Chem. 2002, 41, 4194−4205
Heterobimetallic Bismuth−Transition Metal Salicylate Complexes as
Molecular Precursors for Ferroelectric Materials. Synthesis and
Structure of Bi2M (sal)4(Hsal)4(OR) 4 (M ) Nb, Ta; R ) CH CH ,
2
2
3
i
CH(CH ) ), Bi2Ti3(sal)8(Hsal)2, and Bi2Ti4(O Pr)(sal)10(Hsal) (sal )
3 2
O CC H -2-O; Hsal ) O CC H -2-OH)
2
6 4
2
6 4
John H. Thurston and Kenton H. Whitmire*
Department of Chemistry, Rice UniVersity, 6100 Main Street, Houston, Texas 77005
Received February 18, 2002
The reactions between triphenylbismuth, salicylic acid, and the metal alkoxides M(OCH CH ) (M ) Nb, Ta) or
2
3 5
Ti{OCH(CH ) }4 have been investigated under different reaction conditions and in different stoichiometries. Six
3 2
novel heterobimetallic bismuth alkoxy−carboxylate complexes have been synthesized in good yield as crystalline
solids. These include Bi2M (sal)4(Hsal)4(OR)4 (M ) Nb, Ta; R ) CH CH , CH(CH ) ), Bi2Ti3(sal)8(Hsal)2, and Bi2-
2
2
3
3 2
i
Ti4(OPr)(sal)10(Hsal) (sal ) O CC H -2-O; Hsal ) O CC H -2-OH). The complexes have been characterized
2
6
4
2
6 4
spectroscopically and by single-crystal X-ray diffraction. Compounds of the group V transition metals contain metal
ratios appropriate for precursors of ferroelectric materials. The molecules exhibit excellent solubility in common
organic solvents and good stability against unwanted hydrolysis. The nature of the thermal decomposition of the
complexes has been explored by thermogravimetric analysis and powder X-ray diffraction. We have shown that
the complexes are converted to the corresponding oxide by heating in an oxygen atmosphere at 500 °C. The
mass loss of the complexes, as indicated by thermogravimetric analysis, and the resulting unit cell parameters of
the oxides are consistent with the formation of the desired heterobimetallic oxide. The complexes decomposed to
form the bismuth-rich phases Bi4Ti3O and Bi5Nb3O as well as the expected oxides BiMO (M ) Nb, Ta) and
12
15
4
Bi2Ti4O .
11
Introduction
molecular level.4 Single source precursors are also attractive
because they present a direct route to advanced materials
and because they may require lower temperatures or shorter
calcination periods to be converted to the correct oxide phase
than conventional solid-state methods. Lower preparation
temperatures allow the materials to be deposited on a broader
range of substrates.
A number of bismuth oxide containing materials have
desirable properties, including high Tc superconductivity,5,6
nonlinear optical properties,7 and high oxide ion mobility in
the solid state.8 Of particular significance is a series of
Heterometallic alkoxide or carboxylate complexes are
attractive molecular precursors for bimetallic or multimetallic
oxide materials. These complexes may be amenable for use
in sol-gel synthesis,1 metal-organic chemical vapor deposi-
tion (MOCVD),2 or metal-organic decomposition type syn-
theses of oxides.3 Using discrete, well-defined molecular
precursors may allow control of the metal composition of
these complexes so that they could be used as single-source
precursors for advanced ceramics. A single source precursor
may lead to a higher quality of product because of the
intimate mixing of the different metals achieved at the
(4) Annika, I.; Pohl, M.; Westin, L. G.; Kritikos, M. Chem.sEur. J. 2001,
7, 3438-3445.
(5) Maeda, H.; Tanaka, Y.; Fukutomi, M.; Asano, T. Jpn. J. Appl. Phys.
1988, 27, L209.
* To whom correspondence should be addressed. E-mail: whitmir@
ruf.rice.edu.
(1) Park, Y.; Nagai, M.; Miyayama, M. J. Mater. Sci. 2001, 36, 1261-
(6) Hodge, P.; James, S.; Norman, N.; Orpen, A. J. Chem. Soc., Dalton
Trans. 1998, 4049-4054.
(7) Parola, S.; Papiernik, R.; Hubert-Pfalzgraf, L. G.; Jagner, S.; Hakans-
son, M. J. Chem. Soc., Dalton Trans. 1998, 737-739.
(8) Pell, J. W.; Davis, W. C.; Loye, H. C. z. Inorg. Chem. 1996, 35, 5754.
1269.
(2) Hironaka, K.; Isobe, C.; Moon, B. Jpn. J. Appl. Phys. 2001, 40, 680-
686.
(3) Hou, Y.; Xu, X.; Wang, H. Appl. Phys. Lett. 2001, 78, 1733-1735.
4194 Inorganic Chemistry, Vol. 41, No. 16, 2002
10.1021/ic0255454 CCC: $22.00 © 2002 American Chemical Society
Published on Web 07/18/2002

Molecular Precursors for Ferroelectric Materials
layered oxide materials known as the Aurivillius phases. All
of the Aurivillius phases show ferroelectric behavior and are
therefore potential candidates for ferroelectric random access
memories (FRAM).9 The composition of the Aurivillius
CH2C(CH3)CH2)(THF)][BF4] as well-characterized potential
precursors for advanced ferroelectric materials.
We wish to report here the synthesis and characterization
of six new heterobimetallic complexes of bismuth, Bi2M2-
(sal)4(Hsal)4(OR)4 (1a, M ) Nb, R ) CH(CH3)2; 2a, M )
Ta, R ) CH(CH3)2; 1b, M ) Nb, R ) CH2CH3; 2b, M )
Nb, R ) CH2CH3), Bi2Ti3(sal)8(Hsal)2 (3), and Bi2Ti4(Oi-
Pr)(sal)10(Hsal) (4) (sal ) O2CC6H4-2-O) (Hsal ) O2CC6H4-
2-OH). All of the complexes demonstrate good solubility in
most common organic solvents and greatly improved stability
against hydrolysis compared to the monometallic metal
alkoxides, making these compounds attractive potential
precursors for advanced materials. Powder X-ray diffraction
studies of the complexes reveal that under the conditions
used, the complexes decomposed yielding known bimetallic
oxides. Calcination of 1b and 2b produced the expected
BiMO4 (M ) Nb, Ta) as the major product in both cases,
with small amounts of Bi5Nb3O15 and Bi2O3 being detected
in the two samples, respectively. Interestingly, 3 and 4
decomposed to produce the ferroelectric oxide phase Bi4-
Ti3O12 as the major crystalline phase while the expected Bi2-
Ti4O11 was produced in comparatively small amounts.
phases can be expressed by the general formula (Bi2O2)2+-
2-
(An-1BnO3n+1
)
where A and B can be any metal ion with
the appropriate charge-to-radius ratio.10 Most commonly, A
is Bi, Pb, Sr, Ca, or K, and B is Ti, Zr, Nb, Ta, Mo, W, Fe,
and Co. Of all the potential combinations that have been
created from the general formula, those containing titanium,
niobium, or tantalum have shown the greatest promise for
the development of significant ferroelectric materials. Ex-
amples of these compounds are SrBi2Nb2O9, BaBi2Ta2O9,
and Bi4Ti3O12. These compounds exhibit retention of a high
dielectric constant, low leakage current, and good ferroelec-
tric switching characteristics.7,9,11
The importance of bismuth-containing layered oxide
materials and the lack of well-characterized heterometallic
alkoxide and carboxylate complexes have prompted us to
investigate the synthesis and characterization of complexes
of bismuth with group IV or group V transition metals. There
is relatively little structural information available for het-
erobimetallic alkoxide or carboxylate complexes of bismuth.
Known compounds are limited to several alkali or alkaline-
earth metal species such as KBi(OtBu)4,12 Na2Bi4O2-
(OC6F5)10,13 Bi4Ba4(µ6-O)2(µ3-OEt)8(µ-OEt)4(η2-thd)4,7 and
Na4Bi2(µ6-O)(OC6F5)8(THF)4.14 A few bismuth-transition
metal alkoxide complexes such as [BiCl3OV(OC2H4OMe)3]2,8
BiTi2(µ3-O)(µ-OiPr)4(OiPr)5,15 and Cp2MoBi(µ2-OEt)2(OEt)2-
Cl (Cp ) C5H5)16 have been reported. Bismuth-molybde-
num allyl carbonyl complexes such as [(µ3-CHC(CH3)CH2)-
Mo(CO)2(µ2-OEt)3Bi(OEt)]2 have been characterized.17 Also,
two substituted polyoxometalates of general composition
[NBu4]3[M4Mo(µ5-O)2(µ2-OMe)4(µ2-O)8(NO)(O)4]2Bi (M )
Mo, W) are known.18 Of the compounds reported, [BiCl3-
OV(OC2H4OMe)3]2, Cp2MoBi(µ2-OEt)2(OEt)2Cl, and many
of the allyl-molybdenum complexes contain halide ions that
would likely act as a contaminant of the final oxide material.
The polyoxometalates do not contain the correct stoichiom-
etry of metals to make them useful precursors for the
formation of the ferroelectric Aurivillius phases. This leaves
only BiTi2(µ3-O)(µ-OiPr)4(OiPr)5, Bi4Ba4(µ6-O)2(µ3-OEt)8(µ-
OEt)4(η2-thd)4, Mo2Bi2(CO)4(µ2-OCH2CH3)8(η3-CH2C(CH3)-
CH2)2, and possibly [MoBi(CO)2(µ2-OCH2CH2OCH3)3(η3-
Experimental Section
General Procedures. All manipulations were performed under
an atmosphere of dry nitrogen or argon using standard Schlenk
and glovebox techniques. Solvents were dried over an appropriate
reagent under argon and were distilled immediately prior to use.
Ti(OiPr)4, Nb(OEt)5, Ta(OEt)5, BiPh3, PbPh4 (Strem), and salicylic
acid (Aldrich Chemical Co.) were used as received. Galbraith
Laboratories performed elemental analyses of all of the complexes.
Infrared spectra were recorded on a Nicolet Nexus 670 FT-IR
instrument using attenuated total reflectance (ATR) or on a Nicolet
205 FT-IR as Nujol mulls. NMR data were collected at 298 K on
a Bruker 200, 400, or 500 MHz instrument and were referenced to
the protio impurity in the deuterated solvent (1H) or to tetrameth-
ylsilane (13C) as an internal standard. Chemical shifts are reported
in parts per million (ppm), while coupling constants are given in
hertz (Hz). Coupling constants were not resolved for the aromatic
protons of 1a, 1b, 2a, or 2b. The following abbreviations are
employed in the descriptions of the spectra: s singlet, d doublet, t
triplet, q quartet, h heptet, br broad, Ar aryl. The ipso carbon of
the salicylate ligand was generally not resolved, even with extended
collection times. This can be attributed to the relaxation time of
this quaternary carbon. Powder X-ray diffraction studies were
performed on a Siemens D5000 Diffraktometer, using Cu KR
radiation (λ ) 1.54 Å). Data were collected on 2θ from 10° to 65°
at 0.1° increments with 10 s exposures per frame. Mass spectra of
3 were taken on a Finnigan MAT-95 instrument, using high-
temperature EI (HTEI) techniques. A 100 eV cone voltage was
employed in the experiments. The sample was volatilized by heating
the probe from ambient temperature to 800 °C over the course of
8 min. Data were collected in the range m/z ) 500-2500. Repeated
attempts to study samples 1a, 2a, 1b, 2b, and 4 using DEI, MALDI,
FAB+, ESI, and HTEI techniques did not produce meaningful data.
The solution phase molar masses of complexes 1b and 3 were
determined in THF, using literature procedures.19 Tetraphenyl lead
was used as the reference standard in the solution phase molar mass
(9) Zhou, Q. F.; Chan, H. L. W.; Choy, C. L. J. Non-Cryst. Solids 1999,
254, 106-111.
(10) Sugimoto, W.; Shirata, M.; Sugahara, Y.; Kuroda, K. J. Am. Chem.
Soc. 1999, 121, 11601-11602.
(11) Du, X.; Chen, I. Mater. Res. Soc. Symp. Proc. 1998, 493, 261-266.
(12) Veith, M.; Yu, E.; Huch, V. Chem.sEur. J. 1995, 1, 26.
(13) Whitmire, K. H.; Hoppe, S.; Sydora, O.; Jolas, J. L.; Jones, C. M.
Inorg. Chem. 2000, 39, 85-97.
(14) Jolas, J.; Hoppe, S.; Whitmire, K. Inorg. Chem. 1997, 36, 3335-
3340.
(15) Parola, S.; Papiernik, R.; Hubert-Pfalzgraf, L. G.; Jagner, S.; Hakans-
son, M. J. Chem. Soc., Dalton Trans. 1997, 4631.
(16) Hunger, M.; Limberg, C.; Kircher, P. Angew. Chem., Int. Ed.. 1999,
38, 1105.
(17) Hunger, M.; Limberg, C.; Kircher, P. Organometallics 2000, 19, 1044.
(18) Villanneau, R.; Proust, A.; Robert, F.; Gouzerh, P. J. Chem. Soc.,
Dalton Trans. 1999, 421.
(19) Francis, J. A.; McMahon, C. N.; Bott, S. G.; Barron, A. R.
Organometallics 1999, 18, 4399-4416.
Inorganic Chemistry, Vol. 41, No. 16, 2002 4195

Thurston and Whitmire
experiments. For the TGA/DTA analyses, approximately 5 mg of
the complex to be studied was placed in a platinum pan in the
furnace of a Seiko TGA/DTA 200 instrument. The sample was
heated to 600 °C at a rate of 10 °C/min. The mass loss of each
sample was monitored and compared to that expected for the
formation of the heterobimetallic oxide. Phase changes in the
product heterobimetallic oxide were monitored by DTA.
15 h at room temperature. The product was isolated by removal of
the solvent under reduced pressure, followed by extraction of the
residue with dichloromethane (20 mL) and filtration. The filtrate
was layered with hexane and stored at -20 °C for 2 days, which
resulted in the growth of white needles. These crystals were
collected by filtration and stored under nitrogen. Yield (1b): 0.51
g (54%). Anal. % obsd (% calcd for Bi2Nb2C64H56O28): C, 37.44
(37.44); H, 3.00 (2.75). 1H NMR (C6D6): 0.9 (t, CH3, 3H, JH-H
)
Synthesis. Reactions of BiPh3 with Salicylic Acid. Bismuth
salicylate (Bi{Hsal}3) can be prepared through the stoichiometric
reaction of triphenyl bismuth (0.440 g, 1.0 mmol) with salicylic
acid (0.441 g, 3.0 mmol) in refluxing toluene. The complex formed
in this method is very poorly soluble and precipitates from solution,
leading us to assume that it is oligomeric. The product of this
reaction analyzes for the expected Bi(salH)3 composition: % obsd
(% calcd for BiC21H15O9) C, 40.51 (40.66); H, 2.76 (2.44). IR:
ν(COO) cm-1 1320, 1575, 1599, 1621, 1650; ν(OH) cm-1 3231.
We have found that the use of excess salicylic acid (>7 equiv)
does not lead to precipitation. Optimal conditions for the syntheses
were found to be ca. 20 equiv of salicylic acid per bismuth. In this
case, the resulting bismuth complex is soluble in a wide range of
solvents. Meaningful elemental analyses of that complex have not
been obtained because of contamination by excess salicylic acid.
The complex is most soluble when it is used directly after it has
been prepared. This methodology was employed for the syntheses
of the heterobimetallic complexes discussed in this paper.
4.1 Hz), 3.3 (q, CH2, 2H, JH-H ) 4.1 Hz), 6.4 (s, ArH), 6.6-7.0
(mult, ArH), 7.7 (s, ArH). 13C NMR {1H}(C6D6): 30.6 (CH3), 63.3
(CH2), 118.6 (ArC), 119.7 (ArC), 128.9 (ArC), 131.7 (ArC), 136.9
(COH), 162.8 (CO2). IR ν(COO): 1352s 1572s, 1591s, 1618s.
TGA: % loss (% calcd for formation of BiNbO4) 59% (61%). Yield
(2b): 0.67 g (71%). Anal. % obsd (% calcd for Bi2Ta2C64H56O28):
1
C, 40.18 (40.95); H, 3.28 (3.01). H NMR (C6D6): 0.9 (t, CH3),
3.3 (q, CH2), 6.3 (q), 6.5-7.0 (mult), 7.8 (s). 13C{1H} NMR
(C6D6): 18.6 (CH3), 21.8 (CH2), 118.5 (ArC), 119.7 (ArC), 129.7
(ArC), 131.6 (ArC), 137.0 (COH), 163.1 (CO2). IR: ν (COO)
1352s, 1593s, 1622s, 1660s. TGA: % loss (% calcd for formation
of BiTaO4) 56% (56%).
Bi2Ti4(OiPr)(sal)10(Hsal)‚3HOiPr‚H2O (3). Triphenylbismuth
(0.440 g, 1.0 mmol) and salicylic acid (3.03 g, 22 mmol) were
combined in 20 mL of toluene. The suspension was refluxed for 1
h, and then, the solvent was removed under reduced pressure. The
yellow solid residue was redissolved in 10 mL of a 1 M solution
of water in 2-propanol to give a clear colorless solution. Dropwise
addition of 0.6 mL of Ti(OiPr)4 (2.0 mmol) resulted in a clear
orange-red solution, which was allowed to stand undisturbed for
two weeks. During this time, red crystals deposited in the flask.
The crystals were collected by filtration, washed with 2-propanol
(3 × 10 mL) and pentane (5 mL), and dried in vacuo. Single crystals
suitable for X-ray diffraction were obtained directly from the
reaction mixture. Yield (3): 0.41 g (0.2 mmol, 75%). Anal. % obsd
(% calcd for Bi2Ti4C80H52O34‚3C3H8O‚H2O): C, 45.51 (45.54); H,
Bi2M2(sal)4(Hsal)4(OR)4 (1a, M ) Nb, R ) CH(CH3)2; 2a, M
) Ta, CH(CH3)2). The procedure for the synthesis of these
compounds is similar, so only a general method will be reported
here. Triphenylbismuth (0.440 g, 1.0 mmol) and salicylic acid (3.03
g, 22 mmol) were combined in 20 mL of toluene. The suspension
was refluxed for 1 h, and then, the solvent was removed under
reduced pressure. The yellow solid residue was dissolved in 10
mL of a 1 M solution of water in 2-propanol to give a clear colorless
solution. Dropwise addition of Nb(OEt)5 (0.5 mL, 2.0 mmol)
resulted in a clear yellow solution, which was allowed to stand
undisturbed for two weeks. During this time, yellow crystals
deposited in the flask. The crystals were collected by filtration,
washed with 2-propanol (3 × 10 mL) and pentane (5 mL), and
then dried in vacuo. Single crystals suitable for X-ray diffraction
were obtained directly from the reaction mixture. Yield (1a): 0.58
g (60%). Anal. % obsd (% calcd for Bi2Nb2C68H64O28): C, 42.32
(42.26); H, 3.39 (3.34). 1H NMR (CDCl3): 1.1 (d, CH3, 6H, JH-H
) 4.2 Hz), 3.9 (sp, CH, 1H, JH-H ) 4.2 Hz), 6.5 (br s, ArH), 7.0
(br s, ArH), 7.3 (br s, ArH), 7.8 (br s, ArH), 11.0 (br s, OH). 13C
NMR {1H}(C6D6): -30.6 (CH3), 63.3 (CH), 111.4 (ArC), 118.6
(ArC), 119.7 (ArC), 129.7 (ArC), 131.7 (ArC), 136.9 (COH), 162.8-
(CO2). IR: ν(COO) 1349, 1571, 1591, 1616. TGA: % loss (%
calcd for formation of BiNbO4) 62% (62%). Yield (2a): 0.51 g
(48%). Anal. % obsd (% calcd for Bi2Ta2C68H64O28): C, 38.60
(38.72); H, 3.27 (3.06). 1H NMR (CDCl3): 1.1 (d, CH3, 6H, JH-H
) 6.9 Hz), 3.9 (sp, CH, 1H, JH-H ) 6.9 Hz), 6.9-7.9 (br mult).
13C{1H} NMR (C6D6): -30.4 (CH3), 63.1 (CH), 118.3 (ArC), 119.0
(ArC), 129.5 (ArC), 131.7 (ArC), 137.0 (COH), 162.8 (CO2). IR:
ν(COO) 1352, 1568,1593, 1622. TGA: % loss (% calcd for
formation of BiTaO4) 55% (57%).
1
2.97 (3.29). H NMR (CDCl3): 1.2 (d, CH3, 6H, 6.1 Hz), 4.7
(septet, CH, 1H, 6.1 Hz), 6.2-8.3 (mult, ArH, 1.7 Hz), 9.1 (br s,
OH). 13C{1H} NMR (CDCl3): 25.8 (CH3), 64.9 (CH), 125.7 (ArC),
128.6 (ArC), 129.4 (ArC), 131.3 (ArC), 138.3 (COH), 163.4 (CO2).
HTEIMS: 2169 (M+, 30%), 2113 (A, M+- OiPr, 35%), 2052 (B,
M+- sal, 44%), 1991 (A - sal, 57%), 1929 (B - sal, 80%), 1865
(A - 2sal, 100%), 1801 (B - 2sal, 95%), 1734 (A - 3sal, 70%),
1668 (B - 3sal, 53%), 1600 (A - 4sal, 26%), 1531 (B - 4sal,
23%), 1460 (A - 5sal, 23%). IR: ν(COO) 1391, 1579, 1600, 1604.
TGA: % loss (% calcd for formation of Bi2Ti4O11) 61% (60%).
Bi2Ti3(sal)8(Hsal)2‚3CH2Cl2 (4). Triphenylbismuth (0.440 g, 1.0
mmol) and salicylic acid (0.551 g, 4 mmol) were combined in 20
mL of toluene. The suspension was refluxed for 1 h. During this
time, a yellow precipitate developed. Dropwise addition of 0.3 mL
of Ti(OiPr)4 (1.0 mmol) resulted in an immediate color change to
deep orange and dissolution of all of the solid. The resulting solution
was stirred at room temperature for 15 h, and then, the solvent
was evaporated to dryness under reduced pressure. The orange
residue was extracted with 20 mL of CH2Cl2 and filtered through
Celite. The clear orange filtrate was layered with hexane, and
solvent diffusion was allowed to occur. After 3 days, large orange
crystals had deposited along with a yellow powder. Samples for
X-ray analysis were taken from this solution. The title complex
was isolated by decanting the mother liquor from the crystalline
solid, followed by washing with 2-propanol until the filtrate no
longer became yellow (∼3 × 15 mL). The residue was washed
with pentane (1 × 15 mL) and dried well under vacuum to yield
an orange microcrystalline solid. Yield (4): 0.41 g (75%, based
on bismuth). Anal. % obsd (% calcd for Bi2Ti3C70H42O30‚3CH2-
Bi2M2(sal)4(Hsal)4(OR)4 (1b, M ) Nb, R ) CH2CH3; 2b, M
) Ta, CH2CH3). The procedure for the synthesis of these two
complexes is similar, so only a general method will be described
here. Triphenylbismuth (0.440 g, 1.0 mmol) and salicylic acid
(0.554 g, 4.0 mmol) were refluxed in 20 mL of toluene under argon
for 1 h. After, the reaction period, the yellow suspension was cooled
to room temperature, and tantalum ethoxide (0.26 mL, 1.0 mmol)
was added slowly with stirring. The clear solution was stirred for
4196 Inorganic Chemistry, Vol. 41, No. 16, 2002

Molecular Precursors for Ferroelectric Materials
Table 1. Crystallographic Data for New Compounds
Table 2. Bond Lengths [Å] and Angles [deg] for C68H64Bi2Nb2O28
(1a)
C68H64Bi2-
Nb2O28
C68H64Bi2-
O28Ta2
C80H52Bi2-
O35Ti4
C70H42Bi2-
O30Ti3
Bi(1)-O(43)
Bi(1)-O(13)
Bi(1)-O(33)
Bi(1)-O(32)
Bi(1)-O(42)
Bi(1)-O(23)#1a
2.195(4)
2.242(4)
2.322(4)
2.532(4)
2.621(5)
2.628(4)
Nb(2)-O(51)
Nb(2)-O(61)
Nb(2)-O(11)
Nb(2)-O(21)
Nb(2)-O(22)
Nb(2)-O(12)
1.834(4)
1.851(4)
1.935(4)
1.959(4)
2.069(4)
2.182(4)
fw
1932.97
P1h (No. 2)
1
2109.05
P1h (No. 2)
1
2182.78
P21/n (No. 14)
4
2179.47
P1h (No. 2)
2
space group
Z
cryst syst
a (Å)
b (Å)
triclinic
11.231(2)
12.997(3)
13.615(3)
65.83(3)
72.49(3)
85.07(3)
1727.5(6)
1.858
triclinic
11.361(2)
13.090(3)
13.675(3)
66.06(3)
71.89(3)
84.27(3)
1765.9(6)
1.983
monoclinic
15.143(3)
33.020(7)
19.147(4)
90
108.75(3)
90
9066(3)
1.599
triclinic
13.320(3)
13.654(3)
23.377(5)
82.98(3)
83.68(3)
66.35(3)
3856.5(13)
1.877
O(43)-Bi(1)-O(13)
O(43)-Bi(1)-O(33)
O(13)-Bi(1)-O(33)
O(43)-Bi(1)-O(32)
O(13)-Bi(1)-O(32)
O(33)-Bi(1)-O(32)
O(43)-Bi(1)-O(42)
O(13)-Bi(1)-O(42)
O(33)-Bi(1)-O(42)
O(32)-Bi(1)-O(42)
81.91(16) O(42)-Bi(1)-O(23)#1 74.01(14)
c (Å)
77.14(16) O(51)-Nb(2)-O(61)
81.76(15) O(51)-Nb(2)-O(11)
72.33(17) O(61)-Nb(2)-O(11)
99.5(2)
R (deg)
â (deg)
γ (deg°)
V, Å3
D(calcd),
g‚cm-1
temp (°C)
λ, Mo
97.48(19)
98.60(19)
131.64(14) O(51)-Nb(2)-O(21) 101.9(2)
53.32(14) O(61)-Nb(2)-O(21)
87.33(18)
52.93(15) O(11)-Nb(2)-O(21) 158.49(17)
-50
0.71073
25
0.71073
-50
0.71073
25
0.71073
78.68(15) O(51)-Nb(2)-O(22)
89.79(18)
128.18(15) O(61)-Nb(2)-O(22) 166.74(18)
KR (Å)
113.31(15) O(11)-Nb(2)-O(22)
89.51(17)
81.44(16)
µ (cm-1
R1a
)
54.87
0.0356
0.0911
81.40
0.0367
0.1071
42.86
0.0598
0.1376
51.39
0.0311
0.0835
O(43)-Bi(1)-O(23)#1 124.74(16) O(21)-Nb(2)-O(22)
O(13)-Bi(1)-O(23)#1 72.90(15) O(51)-Nb(2)-O(12) 170.94(17)
wR2b
O(33)-Bi(1)-O(23)#1 142.25(14) O(61)-Nb(2)-O(12)
O(32)-Bi(1)-O(23)#1 154.56(15) O(11)-Nb(2)-O(12)
89.45(18)
79.47(16)
81.69(16)
a Conventional R on Fhkl: ∑|Fo| - |Fc|/∑|Fo|. b Weighted R on |Fhkl| :
2
2
2
2
{∑w(Fo - Fc )2/∑[w(Fo )2]}1/2
.
O(21)-Nb(2)-O(12)
79.94(16) O(22)-Nb(2)-O(12)
a Symmetry transformations used to generate equivalent atoms: #1 -x
+ 2, -y + 2, -z + 1.
1
Cl2): C, 39.87 (40.22); H, 1.34 (2.22). H NMR (CDCl3): 5.8-
8.2 (mult, ArH, JH-H ) 1.7 Hz). 13C{1H} NMR (CDCl3): 116.6
(ArC), 117.7 (ArC), 121.8 (ArC), 122.2 (ArC), 132.5 (ArC), 136.5
(COH), 165.0 (CO2). TGA: % obsd (% calcd for formation of Bi2-
Ti3O9) 63.5% (63.1%). IR: ν(COO) cm-1 1362, 1388, 1529, 1573,
1599, 1622.
employed, the product is soluble only in coordinating
solvents such as THF and methanol. The elemental analysis
of the solid produced in this reaction supports the formulation
as [Bi(O2CC6H4OH)3]n. The solubility of the bismuth sali-
cylate product changes when using excess salicylic acid
during the synthesis. When a 20:1 ratio of acid to BiPh3 is
used, the product displays good solubility not only in
coordinating solvents such as THF and alcohols but also in
hot toluene.
Crystal Structure Determination. Crystals of all compounds
were manipulated under argon using standard inert atmosphere
techniques. Data collection and refinement parameters are given
in Table 1. Complexes 1a, 3, and 4 were studied on a Bruker Smart
1000 diffractometer equipped with a CCD area detector (θmax
)
∼24°). The data were corrected for Lorentz and polarization effects
and for absorption (SADABS).20 Complex 2a was mounted and
studied on a Bruker-Nonius κ diffractometer (θmax ) 33.4°). Several
reflection intensities were monitored throughout data collection,
and no decay of the crystals was detected. All data were corrected
for Lorentz and polarization effects and absorption (empirical).
Heavy atoms were located using direct methods with the SHELXTL
software package.21 All other non-hydrogen atoms were located
by successive Fourier maps and were refined using the full-matrix
least-squares method on F2. Hydrogen atoms were included in
calculated positions and refined using a riding model. All non-
hydrogen atoms in complexes 1a, 2a, and 4 were refined aniso-
tropically. Complex 3 suffered from a systematic disorder of ca.
20 2-propanol molecules per unit cell. Solvent effects were removed
from the data using the program SQUEEZE in the PLATON
software package.22 The structure was refined similarly to the other
molecules using SHELXTL. Only the metal atoms in the structure
were refined anisotropically because of the limited data set. The
weakness of the data is attributable to the massive solvent disorder
and is discussed in more detail in the Results section.
The bismuth salicylate complex produced from the reaction
of 1 equiv of triphenylbismuth with 20 equiv of salicylic
acid reacts with the metal alkoxides Ti{OCH(CH3)2}4 and
M(OCH2CH3)5 (M ) Nb, Ta) to produce heterobimetallic
complexes in good yield under mild conditions. 2-Propanol
was found to be a good solvent for the niobium and tantalum
reactions, and in this solvent, rapid exchange of the ethoxide
ligands for 2-propoxide occurs as anticipated. The complexes
precipitate from the solvent and can be isolated as stable,
crystalline solids. The aggregation and growth of early
transition metal complexes under similar conditions have
been documented.23,24 We have also demonstrated that similar
reactions are possible using stoichiometric amounts of
carboxylic acid in the reaction mixture in the case of
complexes 1b, 2b, and 4. Complexes 1a, 2a, 3, and 4 have
been characterized by single-crystal X-ray diffraction. Se-
lected crystallographic data and bond distances for the data
are presented in Tables 1-5. Diagrams of the complexes,
along with the adopted numbering schemes of the metals,
are presented in Figures 1-6. Complexes 1a and 2a are
isolated from the reaction solutions as yellow or colorless
cubic isomorphous crystals, respectively. The transition metal
atoms in both molecules exist in distorted octahedral
environments, retaining two terminal alkoxide ligands and
Results
Triphenylbismuth reacts with salicylic acid (1:3) in re-
fluxing toluene to produce a stable yellow complex that
precipitates from solution. When this stoichiometry is
(20) Sheldrick, G. SADABS 5.1; University of Go¨ttingen: Go¨ttingen,
Germany, 1997.
(21) Sheldrick, G. SHELXTL 6.1; University of Go¨ttingen: Go¨ttingen,
Germany, 2001.
(22) Spek, A. L. Platon, A Multipurpose Crystallographic Tool; Utrecht
University: Utrecht, The Netherlands, 2001.
(23) Kickelbick, G.; Schubert, U. J. Chem. Soc., Dalton Trans. 1999, 1301-
1305.
(24) Kickelbick, G.; Schubert, U. 1998, 198, 159-161.
Inorganic Chemistry, Vol. 41, No. 16, 2002 4197

Thurston and Whitmire
Table 3. Bond Lengths [Å] and Angles [deg] for C68H64Bi2O28Ta2 (2a)
Table 4. Bond Lengths [Å] and Angles [deg] for C80H52Bi2O35Ti4 (3)
Bi(1)-O(13)
Bi(1)-O(23)
Bi(1)-O(53)
Bi(1)-O(52)
Bi(1)-O(12)
Bi(1)-O(33)#1a
2.204(6)
2.253(5)
2.346(5)
2.548(5)
2.626(7)
2.645(5)
Ta(1)-O(41)
Ta(1)-O(61)
Ta(1)-O(21)
Ta(1)-O(31)
Ta(1)-O(32)
Ta(1)-O(22)
1.839(6)
1.855(6)
1.940(6)
1.964(6)
2.061(5)
2.162(5)
Bi(1)-O(123)
Bi(1)-O(92)
Bi(1)-O(71)
Bi(1)-O(63)
Bi(1)-O(91)
Ti(1)-O(111)
Ti(1)-O(101)
Ti(1)-O(23)
Ti(1)-O(33)
Ti(1)-O(102)
Ti(1)-O(22)
Bi(2)-O(21)
Bi(2)-O(13)
Bi(2)-O(81)
Bi(2)-O(59)
Bi(2)-O(12)
Bi(2)-O(22)
Bi(2)-C(17)
2.195(11)
2.223(10)
2.275(10)
2.438(11)
2.494(10)
1.730(9)
1.878(11)
1.881(10)
2.031(11)
2.058(10)
2.141(10)
2.174(10)
2.245(9)
Ti(2)-O(121)
Ti(2)-O(83)
Ti(2)-O(31)
Ti(2)-O(122)
Ti(2)-O(32)
Ti(2)-O(82)
Ti(3)-O(11)
Ti(3)-O(43)
Ti(3)-O(61)
Ti(3)-O(42)
Ti(3)-O(12)
Ti(3)-O(62)
Ti(4)-O(52)
Ti(4)-O(73)
Ti(4)-O(41)
Ti(4)-O(103)
Ti(4)-O(51)
Ti(4)-O(72)
1.815(10)
1.849(11)
1.881(11)
2.037(11)
2.053(10)
2.069(10)
1.822(10)
1.852(10)
1.853(10)
2.007(9)
O(13)-Bi(1)-O(23)
O(13)-Bi(1)-O(53)
O(23)-Bi(1)-O(53)
O(13)-Bi(1)-O(52)
O(23)-Bi(1)-O(52)
O(53)-Bi(1)-O(52)
O(13)-Bi(1)-O(12)
O(23)-Bi(1)-O(12)
O(53)-Bi(1)-O(12)
O(52)-Bi(1)-O(12)
81.9(2)
78.1(2)
81.0(2)
73.0(2)
O(12)-Bi(1)-O(33)#1 73.56(18)
O(41)-Ta(1)-O(61)
O(41)-Ta(1)-O(21)
O(61)-Ta(1)-O(21)
99.6(3)
97.5(3)
98.2(3)
101.4(3)
88.5(3)
158.6(2)
89.8(2)
167.5(2)
88.7(2)
81.4(2)
171.3(2)
89.0(2)
79.4(2)
82.0(2)
2.022(10)
2.032(10)
1.802(9)
131.01(19) O(41)-Ta(1)-O(31)
53.37(18) O(61)-Ta(1)-O(31)
52.8(2)
78.4(2)
128.6(2)
114.5(2)
O(21)-Ta(1)-O(31)
O(41)-Ta(1)-O(32)
O(61)-Ta(1)-O(32)
O(21)-Ta(1)-O(32)
O(31)-Ta(1)-O(32)
O(41)-Ta(1)-O(22)
2.338(10)
2.382(10)
2.576(10)
2.699(9)
1.827(9)
1.988(10)
2.009(11)
2.031(9)
O(13)-Bi(1)-O(33)#1 124.4(2)
O(23)-Bi(1)-O(33)#1 73.3(2)
2.808(16)
2.048(9)
O(53)-Bi(1)-O(33)#1 141.61(19) O(61)-Ta(1)-O(22)
O(123)-Bi(1)-O(92)
O(123)-Bi(1)-O(71)
O(92)-Bi(1)-O(71)
O(123)-Bi(1)-O(91)
O(92)-Bi(1)-O(91)
O(71)-Bi(1)-O(91)
O(63)-Bi(1)-O(91)
O(123)-Bi(1)-C(97)
O(92)-Bi(1)-C(97)
O(71)-Bi(1)-C(97)
O(63)-Bi(1)-C(97)
O(91)-Bi(1)-C(97)
O(111)-Ti(1)-O(101)
O(111)-Ti(1)-O(23)
O(101)-Ti(1)-O(23)
O(111)-Ti(1)-O(33)
O(101)-Ti(1)-O(33)
O(23)-Ti(1)-O(33)
O(111)-Ti(1)-O(102)
O(101)-Ti(1)-O(102)
O(23)-Ti(1)-O(102)
O(33)-Ti(1)-O(102)
O(111)-Ti(1)-O(22)
O(101)-Ti(1)-O(22)
O(23)-Ti(1)-O(22)
O(33)-Ti(1)-O(22)
O(121)-Ti(2)-O(31)
O(83)-Ti(2)-O(31)
O(121)-Ti(2)-O(122)
O(83)-Ti(2)-O(122)
O(31)-Ti(2)-O(122)
O(121)-Ti(2)-O(32)
O(83)-Ti(2)-O(32)
O(31)-Ti(2)-O(32)
O(122)-Ti(2)-O(32)
O(121)-Ti(2)-O(82)
O(83)-Ti(2)-O(82)
O(31)-Ti(2)-O(82)
O(122)-Ti(2)-O(82)
O(32)-Ti(2)-O(82)
O(11)-Ti(3)-O(43)
O(11)-Ti(3)-O(61)
O(43)-Ti(3)-O(61)
O(11)-Ti(3)-O(42)
O(43)-Ti(3)-O(42)
O(61)-Ti(3)-O(42)
O(11)-Ti(3)-O(12)
O(43)-Ti(3)-O(12)
O(61)-Ti(3)-O(12)
74.5(4) O(123)-Bi(1)-O(63)
76.4(4) O(92)-Bi(1)-O(63)
80.4(3) O(71)-Bi(1)-O(63)
124.7(4) O(102)-Ti(1)-O(22)
54.3(3) O(21)-Bi(2)-O(13)
75.9(4) O(21)-Bi(2)-O(81)
101.2(4) O(13)-Bi(2)-O(81)
100.5(5) O(21)-Bi(2)-O(59)
28.3(4) O(13)-Bi(2)-O(59)
75.6(4) O(81)-Bi(2)-O(59)
92.0(4) O(21)-Bi(2)-O(12)
26.1(4) O(13)-Bi(2)-O(12)
97.9(5) O(81)-Bi(2)-O(12)
99.7(5) O(59)-Bi(2)-O(12)
93.5(5) O(21)-Bi(2)-O(22)
92.4(5) O(13)-Bi(2)-O(22)
164.8(5) O(81)-Bi(2)-O(22)
95.9(5) O(59)-Bi(2)-O(22)
93.9(5) O(12)-Bi(2)-O(22)
85.0(4) O(21)-Bi(2)-C(17)
166.4(4) O(13)-Bi(2)-C(17)
83.1(4) O(81)-Bi(2)-C(17)
174.3(5) O(59)-Bi(2)-C(17)
87.3(4) O(12)-Bi(2)-C(17)
82.2(4) O(22)-Bi(2)-C(17)
82.1(4) O(121)-Ti(2)-O(83)
98.4(5) O(42)-Ti(3)-O(12)
98.6(5) O(11)-Ti(3)-O(62)
84.7(4) O(43)-Ti(3)-O(62)
94.1(5) O(61)-Ti(3)-O(62)
166.6(5) O(42)-Ti(3)-O(62)
96.9(5) O(12)-Ti(3)-O(62)
167.4(5) O(52)-Ti(4)-O(73)
84.1(4) O(52)-Ti(4)-O(41)
82.5(4) O(73)-Ti(4)-O(41)
88.1(4)
80.1(3)
157.8(3)
84.2(4)
76.1(4)
84.6(4)
79.4(3)
91.8(3)
70.1(3)
149.2(3)
129.3(3)
54.1(3)
78.4(3)
80.5(3)
52.4(3)
116.5(3)
122.5(3)
76.6(3)
157.1(3)
102.6(4)
26.6(4)
80.9(4)
70.1(4)
27.9(3)
137.6(4)
94.9(6)
79.6(4)
165.3(5)
99.8(4)
84.2(4)
81.1(4)
82.3(4)
99.8(4)
90.6(4)
95.6(4)
O(52)-Bi(1)-O(33)#1 154.6(2)
O(31)-Ta(1)-O(22) 80.4(2)
O(21)-Ta(1)-O(22)
O(32)-Ta(1)-O(22)
a Symmetry transformations used to generate equivalent atoms: #1 -x
+ 1, -y + 1, -z + 1.
two chelating salicylate ions. The niobium-oxygen bond
lengths range from 1.834(4) to 2.182(4) Å in 1a. The
tantalum-oxygen bond distances range from 1.839(6) to
2.162(5) Å in 2a. The shortest bonds, which are associated
with the alkoxide ligands in both cases, are those that are
approximately trans to the carboxylate group of a salicylate
ligand. These carboxylate-metal bonds are the longest M-O
distances in the molecule.
The bismuth atoms are also six-coordinate, but the
geometry at the metal center is severely distorted from an
octahedral or trigonal prismatic environment (Figure 7). The
bismuth atoms possess two terminal salicylate ligands that
chelate the metal through the carboxylate functionality. A
third salicylate ion bridges between the bismuth atom and
the transition metal. The coordination sphere of the metal is
completed by dative interactions to a fourth carboxylate
ligand.
In both 1a and 2a, the aromatic ring of each of the bridging
ligands is bent toward one of the bismuth atoms suggesting
a weak π-interaction (Figures 1 and 7). The distance between
the ring centroid and the bismuth atom is 3.393 Å in the
case of 1a and is 3.410 Å in the case of 2a. The bismuth-
oxygen bond lengths in the complexes vary over a large
range. In 1a, the bond distances range from 2.195(4) to
2.628(4) Å. In 2a, the bond lengths range from 2.204(6) to
2.645(5) Å. There appears to be a rough correlation in each
case between the bond lengths so that a long metal-oxygen
bond is located approximately trans to a short metal-oxygen
bond. In addition, the chelating salicylate ligands that are
bound only to the bismuth atom each contain one long and
one short metal oxygen bond. The shortest metal oxygen
bond is roughly trans to the coordinated aromatic ring.
164.6(4) O(52)-Ti(4)-O(103) 100.2(4)
84.9(5) O(73)-Ti(4)-O(103) 89.3(4)
96.8(4) O(41)-Ti(4)-O(103) 167.2(4)
80.0(4) O(52)-Ti(4)-O(51)
82.6(4) O(73)-Ti(4)-O(51)
94.8(5) O(41)-Ti(4)-O(51)
95.0(5) O(103)-Ti(4)-O(51)
101.0(5) O(52)-Ti(4)-O(72)
98.2(5) O(73)-Ti(4)-O(72)
85.5(4) O(41)-Ti(4)-O(72)
164.8(5) O(103)-Ti(4)-O(72)
83.1(4) O(51)-Ti(4)-O(72)
164.5(4) Ti(3)-O(12)-Bi(2)
94.5(4) Ti(1)-O(22)-Bi(2)
86.0(4)
170.4(5)
91.9(4)
82.2(4)
172.6(4)
85.2(4)
83.4(4)
85.2(4)
89.8(4)
141.5(4)
144.5(4)
range between 1.730(9) and 2.141(10) Å with the Bi-O bond
distances ranging between 2.175(10) and 2.808(16) Å. The
bismuth atoms are found to be five- or six-coordinate (Figure
8). The metal oxygen bond lengths compare favorably to
those in similar molecules.25
Compound 3 is isolated as deep red prisms. The molecule
was found to be asymmetric, containing six metal atoms in
a variety of coordination environments. As in 1a and 2a,
the transition metals in 3 are six-coordinate and exist in
distorted octahedral environments. The Ti-O bond lengths
4198 Inorganic Chemistry, Vol. 41, No. 16, 2002

Molecular Precursors for Ferroelectric Materials
Table 5. Bond Lengths [Å] and Angles [deg] for C70H42Bi2O30Ti3 (4)
Bi(1)-O(23)
Bi(1)-O(43)
Bi(1)-O(32)
Bi(1)-O(73)
Bi(1)-O(42)
Bi(1)-O(22)
Bi(1)-O(33)
Bi(2)-O(93)
Bi(2)-O(103)
Bi(2)-O(83)
Bi(2)-O(53)
Bi(2)-O(92)
Ti(3)-O(71)
Ti(3)-O(81)
Ti(3)-O(63)
2.238(4)
2.309(4)
2.321(4)
2.417(4)
2.476(4)
2.607(4)
2.623(4)
2.200(4)
2.209(4)
2.344(4)
2.444(4)
2.496(4)
1.816(4)
1.845(4)
1.985(4)
Ti(3)-O(13)
Ti(3)-O(72)
Ti(3)-O(82)
Ti(4)-O(51)
Ti(4)-O(61)
Ti(4)-O(41)
Ti(4)-O(52)
Ti(4)-O(62)
Ti(4)-O(42)
Ti(5)-O(11)
Ti(5)-O(101)
Ti(5)-O(21)
Ti(5)-O(12)
Ti(5)-O(22)
Ti(5)-O(102)
2.008(4)
2.019(4)
2.046(4)
1.838(4)
1.844(4)
1.865(4)
2.018(4)
2.054(4)
2.076(4)
1.832(4)
1.848(4)
1.862(4)
2.045(4)
2.083(4)
2.093(4)
O(23)-Bi(1)-O(43)
O(23)-Bi(1)-O(32)
O(43)-Bi(1)-O(32)
O(23)-Bi(1)-O(73)
O(43)-Bi(1)-O(73)
82.35(14) O(23)-Bi(1)-O(42)
81.96(15) O(43)-Bi(1)-O(42)
77.24(14) O(32)-Bi(1)-O(42)
80.96(15) O(73)-Bi(1)-O(42)
71.56(14) O(23)-Bi(1)-O(22)
135.82(13)
53.74(12)
83.62(14)
88.52(15)
52.90(13)
135.25(12)
88.48(17)
169.27(16)
86.51(16)
173.49(17)
90.25(16)
87.16(16)
175.30(16)
85.97(16)
84.83(16)
84.70(16)
88.82(15)
99.60(17)
94.84(18)
97.14(18)
84.34(17)
99.96(18)
162.78(16)
165.00(18)
84.32(16)
99.05(17)
80.71(16)
95.31(17)
165.08(16)
99.37(17)
82.15(16)
79.67(15)
98.99(17)
O(32)-Bi(1)-O(73) 146.04(14) O(43)-Bi(1)-O(22)
Figure 1. ORTEP representation of 1a. Thermal ellipsoids are drawn at
the 50% probability level. The π-interactions between bismuth and the
salicylate ring are represented by broken lines.
O(32)-Bi(1)-O(22)
O(73)-Bi(1)-O(22)
95.09(14) O(81)-Ti(3)-O(13)
97.64(15) O(63)-Ti(3)-O(13)
O(42)-Bi(1)-O(22) 170.46(12) O(71)-Ti(3)-O(72)
O(23)-Bi(1)-O(33) 104.09(15) O(81)-Ti(3)-O(72)
O(43)-Bi(1)-O(33) 126.39(13) O(63)-Ti(3)-O(72)
O(32)-Bi(1)-O(33)
52.14(14) O(13)-Ti(3)-O(72)
O(73)-Bi(1)-O(33) 161.55(14) O(71)-Ti(3)-O(82)
O(42)-Bi(1)-O(33)
O(22)-Bi(1)-O(33)
99.25(13) O(81)-Ti(3)-O(82)
72.77(13) O(63)-Ti(3)-O(82)
O(93)-Bi(2)-O(103) 79.49(15) O(13)-Ti(3)-O(82)
O(93)-Bi(2)-O(83)
77.82(14) O(72)-Ti(3)-O(82)
O(103)-Bi(2)-O(83) 82.18(14) O(51)-Ti(4)-O(61)
O(93)-Bi(2)-O(53)
74.82(15) O(51)-Ti(4)-O(41)
O(103)-Bi(2)-O(53) 89.00(15) O(61)-Ti(4)-O(41)
O(83)-Bi(2)-O(53) 152.36(14) O(51)-Ti(4)-O(52)
O(93)-Bi(2)-O(92)
55.20(15) O(61)-Ti(4)-O(52)
O(103)-Bi(2)-O(92) 134.14(14) O(41)-Ti(4)-O(52)
O(83)-Bi(2)-O(92)
O(53)-Bi(2)-O(92)
O(71)-Ti(3)-O(81)
O(71)-Ti(3)-O(63)
O(81)-Ti(3)-O(63)
O(71)-Ti(3)-O(13)
O(41)-Ti(4)-O(42)
O(52)-Ti(4)-O(42)
O(62)-Ti(4)-O(42)
81.79(15) O(51)-Ti(4)-O(62)
85.85(15) O(61)-Ti(4)-O(62)
98.72(17) O(41)-Ti(4)-O(62)
94.62(18) O(52)-Ti(4)-O(62)
93.16(17) O(51)-Ti(4)-O(42)
95.61(18) O(61)-Ti(4)-O(42)
81.73(16) O(101)-Ti(5)-O(22)
81.22(15) O(21)-Ti(5)-O(22)
81.19(15) O(12)-Ti(5)-O(22)
Figure 2. ORTEP representation of 2a. Thermal ellipsoids are drawn at
the 50% probability level. The π-interactions between bismuth and the
salicylate ring are represented by broken lines.
O(11)-Ti(5)-O(101) 98.01(19) O(11)-Ti(5)-O(102)
O(11)-Ti(5)-O(21) 95.37(18) O(101)-Ti(5)-O(102) 83.00(16)
O(101)-Ti(5)-O(21) 94.89(18) O(21)-Ti(5)-O(102) 165.64(17)
O(11)-Ti(5)-O(12)
83.90(17) O(12)-Ti(5)-O(102)
81.05(15)
84.18(15)
139.49(17)
135.64(18)
O(101)-Ti(5)-O(12) 164.04(17) O(22)-Ti(5)-O(102)
O(21)-Ti(5)-O(12) 100.73(17) Ti(5)-O(22)-Bi(1)
O(11)-Ti(5)-O(22) 162.59(18) Ti(4)-O(42)-Bi(1)
Solid samples of 3 were found to contain solvent located
both in the lattice and in the interior of the molecule itself.
Interestingly, a single molecule of water was found to occupy
the center of the molecule of 3. In the crystal lattice, the
water molecule is four-coordinate and exhibits a distorted
tetrahedral geometry. The water molecule is presumably
weakly bound to the bismuth atoms (dBi(1)-O(1) ) 2.783(9)
Å, dBi(2)-O(1) ) 3.004(9) Å) through lone pair electrons and
also interacts with two salicylate ligand oxygen atoms (O(32)
and O(102)) through hydrogen bond interactions (dO(32)-O(1)
) 2.757(14) Å, dO(102)-O(1) ) 2.83 Å). A third possible, but
weaker, interaction was also detected between the water
Figure 3. ORTEP representation of 3. Thermal ellipsoids have been drawn
at the 30% probability level for clarity.
molecule and O(72) (dO(72)-O(1) ) 2.92 Å). These bond
distances are comparable with distances that have been
reported for other hydrates, which have been reported to
range from 2.6 to 2.9 Å in length.26
(25) Asato, E.; Katsura, K.; Arakaki, T.; Mikuriya, M.; Kotera, T. Chem.
Lett. 1994, 2123-2126.
Inorganic Chemistry, Vol. 41, No. 16, 2002 4199

Thurston and Whitmire
Figure 6. Winray representation of 4.42 Metal atoms have been emphasized
to illustrate the connectivity of the complex. Hydrogen atoms have been
omitted for clarity. Color scheme: orange, bismuth; red, titanium; blue,
oxygen; yellow, carbon.
Figure 4. Winray representation of 3.42 Metal atoms have been emphasized
to illustrate the connectivity of the complex. Hydrogen atoms have been
omitted for clarity. Color scheme: orange, bismuth; red, titanium; blue,
oxygen; yellow, carbon.
Figure 7. Representation of the coordination environment of the bismuth
atoms in 1a and 2a.
Figure 5. ORTEP representation of 4. Thermal ellipsoids are drawn at
the 30% probability level for clarity.
The assignment of the oxygen atom as water was based
on the long bond distances of the atom to the metal center
and charge balance requirements of the entire complex.
Hydrated bismuth atoms have recently been characterized
as trifluoromethanesulfonate salts.27,28 In these samples, the
Bi-Owater bond lengths lie in the range 2.448-2.577 Å. This
is considerably longer than the reported bond distances for
µ2 bridging oxo ligands, which lie in the range 2.023-2.123
Å.29 The long Bi-Owater bond in 3 can be rationalized as
rising from bridging orientation of the water molecule. This
orientation would be expected to give rise to longer bond
distances, compared to the terminal orientation found in the
Figure 8. Representation of the bismuth coordination environment in 3.
hydrated bismuth triflate structure. The charge balance
requirements of the metal centers in the complex are also
fully satisfied by the salicylate and alkoxide ligands alone.
This further suggests that the identity of the oxygen atom in
the structure is water, which would allow the complex to
remain neutral. Spectroscopic characterization of the water
(26) Wells, A. F. Structural Inorganic Chemistry, 5th ed.; Oxford Press:
Oxford, 1984.
(27) Frank, W.; Reiss, G. J.; Schneider, J. Angew. Chem., Int. Ed. Engl.
1995, 34, 2416-2417.
(28) Na¨slund, J.; Persson, I.; Sandstro¨m, M. Inorg. Chem. 2000, 39, 4012-
4021.
1
by H NMR and FTIR was inconclusive because of the
(29) Hassan, A.; Breeze, S.; Courtenay, S.; Deslippe, C.; Wang, S.
Organometallics 1996, 15, 5613-5621.
presence of residual hydroxyl groups in the salicylate ligands.
4200 Inorganic Chemistry, Vol. 41, No. 16, 2002

Molecular Precursors for Ferroelectric Materials
Figure 9. Ball-and-stick packing diagram of 3 detailing the void area found in the crystal lattice.
Modeling of the crystallographic data of 3 using the
program PLATON reveals that the solid-state structure
contains void area equal to 2020 Å3, or 22% of the unit cell
(Figure 9). Disordered solvent molecules that would con-
tribute to a fairly high residual electron density that is
observed in the structure presumably occupy this void area.
Assuming that a single molecule of 2-propanol has a volume
of approximately 110 Å3, then it appears that as many as 20
molecules of solvent may be located in the unit cell. Although
many of the included molecules are lost during the isolation
of the material, it was possible to confirm the presence and
identity of the lattice solvent through a variety of techniques.
The elemental analysis of 3 agreed with the inclusion of
approximately three molecules of 2-propanol in the crystal
lattice. Further,because the molecules assayed consistently
high for carbon and hydrogen, it was concluded that the
lattice solvent was not exclusively water. The results of
elemental analysis were confirmed through careful inter-
rogation of the samples by proton NMR. Fresh samples of
3 demonstrate two distinct methyne resonances at δ ) 3.98
and 4.76 ppm. The two methyne peaks integrate to ap-
proximately 3.3:1 and can be correlated to the free lattice
solvent and the single 2-propyl group that is located on 3,
respectively. The assignment of the peaks is supported by
examining the NMR spectra of samples of 3 that have either
been aged for several months or dried under vacuum (10-7
Torr) for several days. In these cases, the resonance at 4.76
ppm is unchanged while that at 3.98 is reduced or absent.
This low field shift of the methyne proton in the 2-propoxide
ligands coordinated to titanium relative to the free alcohol
has been observed for other complexes as well.30 The NMR
analysis is supported by TGA analysis, which demonstrates
a quick loss of volatile material comprising approximately
9.7% of the mass of the sample (calculated for three
molecules of 2-propanol: 9%). It is difficult to conclude from
these experiments if the coordinated water is also lost from
the lattice, as the calculated change in mass for loss of the
single water molecule is very small.
The mass spectrum of 3 displays the parent ion at the
expected location. Isotopic matching of the observed spec-
trum supports the assignment of the parent ion peak. In
addition to the parent ion, the spectrum of 3 demonstrates
12 fragments corresponding to the successive loss of sali-
cylate ions or 2-propoxide ions from the parent complex.
Three peaks heavier than the parent ion were also observed
in the spectrum. These can be assigned to products resulting
from exchange reactions or rearrangements of the parent ion.
Compound 4 was synthesized by allowing Ti{OCH-
(CH3)2}4 to react with the product of the reaction of triphenyl
bismuth and 4 equiv of salicylic acid in toluene at room
temperature. These are similar reaction conditions to what
was employed for the syntheses of 1b and 2b. Careful
recrystallization of the product mixture led to the growth of
deep orange prisms of 4 along with the deposition of a yellow
powder. The crystals were isolated and characterized both
crystallographically and spectroscopically.
The overall structure of 4 was found to resemble that of
3, but it contained one less titanium atom, one less salicylate
ion, and no residual alkoxide ligands. The Ti-O bond lengths
in 4 range from 1.816(4) to 2.093(4) Å. This agrees well
with what is observed in 3. The geometry of the transition
metal centers, including bond lengths and angles, is similar
to what has been observed in the structure of 3. Compound
4 exhibits Bi-O bond distances that range between 2.200(4)
and 2.623(4) Å. Like 3, the two bismuth centers in the
compound exhibit different coordination numbers (Figure
10). In this case, one of the bismuth atoms is five-coordinate,
and the other, seven-coordinate. Under the conditions that
were used in the synthesis and purification of 4, the molecule
cocrystallizes with three solvent molecules of dichlo-
romethane. The solvent molecules were ordered in the crystal
lattice and refined well.
The thermal decompositions of 1a-4 were investigated
by TGA. Compound 1a begins to decompose at 200 °C with
a mass loss of 22%. This is followed by two additional mass
losses of 27% and 16% at approximately 280 °C and 385
°C, respectively. Compound 2a is very similar with its
decomposition beginning at 210 °C with a mass loss of 14%.
This is followed by two additional mass losses of 26% and
(30) Motoyama, Y.; Tanaka, M.; Mikami, K. Inorg. Chim. Acta 1997, 256,
161-163.
Inorganic Chemistry, Vol. 41, No. 16, 2002 4201

Thurston and Whitmire
loss required toconvert 4‚3CH2Cl2 to the metal oxide
(calcd 75%, obsd 76%). The final oxide is obtained at 410
°C.
The nature of the oxide that is obtained from each of
the complexes was probed by powder X-ray diffraction.
Samples of heterobimetallic complexes were converted to
the crystalline materials by calcining at temperatures in
excess of 700 °C for 3 h in air. The product was slowly
cooled to room temperature to produce a white or pale yellow
crystalline solid in all cases. The product was ground to
produce a finely dispersed powder that was used in the
analysis. It was observed that calcining the samples at lower
temperatures than those used in these experiments produced
only amorphous material. The identity of the product
materials as the heterobimetallic oxides was confirmed by
comparing the powder diffraction patterns to accepted
standards, using the computer program JADE.31 It has been
demonstrated that 1a and 2a decompose with heating at 700
°C to produce the orthorhombic phase (the low temperature
form) of BiMO4 (M ) Nb, Ta), as the major crystalline
phase. Samples of 1a were also found to contain the bismuth-
rich heterobimetallic oxide Bi5Nb3O15. The oxide of 2a was
found to contain Bi2O3. No crystalline bismuth-rich hetero-
bimetallic oxide was detected in the sample. Complex 3
decomposed to form the bismuth-rich oxide Bi4Ti3O12 as the
major crystalline phase. The expected product Bi2Ti4O11 was
only detected in ca. 30% relative abundance to the bismuth-
rich phase. Complex 4 was found to react similarly on
heating to produce Bi4Ti3O12 and Bi2Ti4O11 in a ca. 90:10
ratio.
Figure 10. Representation of the bismuth coordination environment in 4.
12% at 305 °C and 397 °C, respectively. Mass losses of 62%
and 57% are required for the conversion of 1a and 2a to
metal oxide, respectively. This agrees with the total observed
losses (62% and 55%, respectively). The thermograms of
compounds 1b and 2b are similar to those of 1a and 2a,
asexpected. The decomposition of 1b and 2b occurs at
slightly lower temperature than that of 1a and 2a. Compound
1b demonstrates three successive mass losses at 110, 278,
and 401 °C, totaling 59% of the total sample weight. This
corresponds well with the calculated value of 61% mass loss
for the conversion of 1b to BiNbO4. Similarly, 2b demon-
strates three distinct mass losses at 124, 302, and 365 °C
totaling 56% of the sample, which agrees exactly with the
calculated mass loss required to convert the sample to metal
oxide (56%).
As mentioned earlier, the thermogram of 3 shows a quick
mass loss of 9.7% starting at approximately 35 °C, probably
correlating to the loss of solvent from the crystal lattice. This
has been confirmed experimentally by drying a sample of
the cluster under dynamic vacuum for 1 week. In this case,
the sample lost approximately 12% of its original mass,
which, within the limits of experimental error, agrees well
with the observations of single-crystal X-ray analysis and
1H NMR. Decomposition of 3 begins at approximately 270
°C with a mass loss of 42%. This is followed by a second
mass loss at 390 °C of 19%. A phase change in the material
is observed at 427 °C. The final oxide is obtained at 500
°C. Conversion of 3 completely to metal oxides requires the
60% mass loss from the sample. The final two steps of the
thermal decomposition of 3 represent a 61% mass loss, which
agrees with the expected value.
Solution phase measurements of the molar masses of
complexes 1b and 3 in THF solution gave results of 1227
and 1960 g‚mol-1, respectively. Calculated masses for the
complexes in the solid state are 1876.61 and 2152.63
g‚mol-1, respectively. These results are consistent with the
conjecture that 3 dissolves unchanged but that 1b dissociates
into a solvated monomer (cf. the molar mass of monomer‚
2THF is 1190 g‚mol-1). While there is some uncertainty
associated with the molar mass determined in this manner,
1
when these data are coupled with the H NMR results, it
creates a convincing case that molecules 1a and 1b dissociate
in solution (Figure 11).
Discussion
The use of bifunctional ligands such as salicylic acid offers
a direct “one-pot” synthesis for heterobimetallic complexes
containing bismuth. The fact that the solubility of the
bismuth-salicylate complex changes as the amount of
salicylic acid in the reaction mixture is altered implies that
there is an interaction of the excess salicylic acid with the
complex. It seems likely that the complex that is formed from
an exact 1:3 ratio of bismuth to salicylic acid is oligomeric
or polymeric, existing as [Bi(salH)3]n (salicylateH ) O2-
CC6H4-2-OH). Coordination polymerization is common for
The thermal decomposition of 4 is similar to 3. The
molecule decomposes in two steps. Loss of lattice solvent
from the sample occurs between 37 °C and 152 °C,
represents a loss of 12% of the sample, and is consis-
tent with approximately three molecules of dichlorome-
thane per molecule of complex (ca. 12%). This is in
agreement with the results of the X-ray experiments and
elemental analysis. Decomposition of the complex occurs
in a single step beginning at 278 °C and reflects a 64% loss
of mass from the sample. This total mass loss observed for
the sample is in good agreement with the calculated mass
(31) JADE, XRD Pattern-Processing for the PC 2.1; Materials Data, Inc.:
Livermore, CA, 1994.
4202 Inorganic Chemistry, Vol. 41, No. 16, 2002

Molecular Precursors for Ferroelectric Materials
Figure 11. Dissociation of 1a into a solvated monomer.
Lewis acidic bismuth centers, which are often observed with
high coordination numbers, and elemental analysis of the
complex supports this formulation.14,32
phenoxycomplexes of bismuth such as Bi(OC6H3-2,6-
(CH3)2)3 and [Bi(OC6F5)3(µ6-C7H8)]2 are yellow.13,33,34
Bismuth (III) centers with electronegative ligands are
strongly Lewis acidic, so it is reasonable to conclude that
the metal is able to coordinate oxygen atoms of adjacent
complexes resulting in the oligomerization or polymerization.
It appears that excess salicylic acid disrupts these additional
interactions, probably through the formation of salicylate
adducts such as Bi(salH)3(salH2). This explains the higher
solubility of the product in the presence of excess salicylic
acid. Investigations into the nature of these complexes, which
are difficult to isolate free from excess salicyclic acid, are
ongoing.
Drawings I-VI illustrate possible coordination modes
of the salicylate ligand with bismuth or with bismuth and
a transition metal. Of these, the terminal mode IV has not
been observed in this class of complexes. The structural
observations for 1a, 2a, 3, and 4 suggest that bismuth
prefers to bond through the carboxylate functionality of the
salicylate ligand (I, V, VI), while transition metals associate
with both the hydoxyl and carboxylate functional groups
to form a six-membered chelate ring (II). However, this
does not rule out the possibility of chelating salicylate ligands
that are associated with bismuth through both functional
groups, particularly in absence of other metal species. The
fact that [Bi(O2CC6H4OH)3]n is yellow may indicate that the
hydroxyl groups of the salicylate ligands are interacting with
the metal center (III). This theory is supported by IR studies
of the complex, which demonstrate a ν(OH) peak at 3231
cm-1. This is at a higher frequency than the free acid (ν-
(OH) ) 3059 cm-1), suggesting that there is interaction
between the phenolic oxygen and the bismuth atom. This
conjecture is also supported by the fact that reactions carried
out using triphenyl bismuth and benzoic or methacrylic acid
produce only a white product. Furthermore, many known
Given the high oxophilicity of the group IV and V
transition metals, one of the most interesting outcomes of
the syntheses is the noticeable lack of oxo ligands and the
relative stability of residual alkoxide moieties in 1-4. This
is particularly surprising for complexes 1-3 because of the
fact that the syntheses of these compounds were performed
in the presence of excess carboxylic acid. Under these
conditions, it is known that the unreacted acid can undergo
esterification with liberated alcohol or with the solvent
yielding water as a byproduct and leading to the formation
of oxo or hydroxo ligands. However, these complexes appear
to exhibit exceptional stability toward hydrolysis. Compound
3 was observed to cocrystallize with water with no apparent
(33) Jones, C. M.; Burkhart, M. D.; Bachman, R. E.; Serra, D. L.; Hwu,
S.-J.; Whitmire, K. H. Inorg. Chem. 1993, 32, 5136.
(34) Evans, W. J.; Hain, J. H.; Ziller, J. W. Chem. Commun. 1989, 1628.
(32) Asato, E.; Katsura, K.; Mikuriya, M.; Fuji, T.; Reediijk, J. Inorg. Chem.
1993, 32, 5322-29.
Inorganic Chemistry, Vol. 41, No. 16, 2002 4203

Thurston and Whitmire
hydrolysis or decomposition. Attempts to modify 3 by
addition of /2-1 mmol of water/mmol of titanium during
various stages of the synthesis still gave 3 as the only isolable
product.
magnetically inequivalent salicylate ions. This strongly
suggests that 1a, 2a, 1b, and 2b are able to undergo ligand
exchange processes at a rate that is fast on the NMR time
scale. Attempts to slow the rate of exchange to the point
that individual salicylate environments could be resolved
through cooling of the sample were not successful.
1
The π-interaction of the aromatic ring of the salicylate
ligand with the bismuth center in 1a and 2a (Figure 7)
suggests that the bismuth atom in these complexes is
relatively electron deficient. This type of interaction has been
documented in other complexes that contain bismuth coor-
dinated to electronegative ligands, such as petafluorophen-
oxide,13,33,35 trifluoroacetate,36 and chloride.37-39 The distances
between the metal and the aromatic ring compare well with
other compounds that have been reported by our group,
namely Bi4(µ4-O)(µ2-OC6F5)6{µ3-OBi(µ-OC6F5)3}2(µ6-C6H5-
CH3) and [Bi(µ-OC6F5)(OC6F5)2(µ6-C6H5CH3)]2 which have
bismuth-ring centroid distances of 3.074(2) and 2.968(2)
Å, respectively.16 The bismuth-aromatic ring centroid
distances in these molecules are shorter than in 1a or 2a,
suggesting that the metal center in 1a and 2a is less
electropositive than it is in the pentafluorophenoxy com-
plexes, as might be expected.
The IR spectra of complexes 1a-4 are complex and
difficult to interpret. All of the complexes demonstrate a
strong peak in the region of 3200-3600 cm-1, which
correlates to the O-H stretch of the singly deprotonated
salicylate ligands in the structure. The coordinated water
molecule observed in the solid-state structure of 3 should
also produce a distinct peak in this region. However, this
peak overlaps with the stronger phenolic O-H stretch and
was consequently unobservable. The antisymmetric carbonyl
stretch of free carboxylic acids is known to occur in the range
1800-1550 cm-1, while the symmetric carbonyl stretch is
known to lie in the range approximately 1450-1370 cm-1.40
For free salicylic acid, the νas(COO) and νs(COO) are
observed to be 1662 and 1441 cm-1, respectively. The
interactions of the ligand with the metal centers in complexes
discussed in this paper were probed through monitoring the
shift in frequency of the νas(COO) and νs(COO) peaks of
the free acid. For all of the complexes, the νas(COO) peaks
were found to be significantly shifted toward 1550 cm-1.
This has been noted to be indicative of ionization of the
carboxyl group and supports the results of the single crystal
X-ray studies in indicating that no protons are associated
with the carboxylic functional group of the ligands. The
symmetric carbonyl stretch was found to have shifted to
approximately 1350 cm-1 for the complexes. This peak has
been used to provide structural information regarding the
mode of ligand association between one or more metal
centers in simple cases.41 In our complexes, it is difficult or
impossible to make accurate assignments of specific bonding
modes because of the complexity of the structures and
resultant spectra; however, the range of the frequencies of
the peaks that are observed for these complexes is suggestive
of either bidentate (I) or bidentate/bridging coordination (V
and VI) by the carboxylate ligand. This observation, like
that for the antisymmetric carbonyl peak, agrees with the
findings of the single-crystal structure analyses.
Molar mass studies undertaken in coordinating solvents
indicate that 3 dissolves unchanged, while 1b demonstrates
a significantly different molar mass in solution than in the
solid state. These calculations can be correlated with the
observed solid-state structures of the complexes. Compound
3 is assembled so that all of the metal centers in the complex
are associated through a series of covalent bonds. These
bonds provide the stabilization needed to protect the integrity
of the complex when the molecule is dissolved. In contrast,
1b is observed to be a dimer in the solid state, in which the
two subunits of the complex are associated through dative
interactions. These bonding modes are much weaker than
what is observed in 3, and solvation results in disruption of
the dimeric structure. The monomers, in this case, appear to
be stabilized though coordination interactions of several
solvent molecules with the metal centers and are thought to
exist in dynamic equilibrium with the dimeric complex.
Nuclear magnetic resonance studies of 1a-4 were not
informative either because of the low symmetry of the
complex and resultant large number of peaks (3, 4) or the
fact that apparent rapid ligand exchange of the salicylate ions
produced very broad peaks in the aromatic region (1a, 2a,
1b, 2b). In all cases, the alkoxide ligand resonances were
clearly resolved. The protons adjacent to the alkoxide oxygen
frequently demonstrated a significant downfield shift upon
coordination, which is consistent with metal coordination.30
All of the compounds were easily converted to metal
oxides when the samples were heated in air. TGA analysis
indicates that the complexes undergo complete thermal
decomposition on heating to 600 °C. Thermal decomposition
studies carried out under nitrogen gave similar results. This
suggests that all of the oxygen atoms necessary for the
production of the oxide may be obtained from the initial
complex. The bimetallic oxides obtained at these tempera-
tures are amorphous. Heating the molecular samples to
greater than 700 °C produces crystalline samples of the
bimetallic oxides. Under the conditions that were explored
1
It is significant that the H NMR spectra of 1a, 2a, 1b,
and 2b demonstrate a single salicylate environment, whereas
the spectra of 3 and 4 show the presence of numerous
(35) Jones, C. M.; Burkhart, M. D.; Whitmire, K. H. Angew. Chem., Int.
Ed. Engl. 1992, 31, 451.
(36) Frank, W.; Reiland, V.; Reiss, G. J. Angew. Chem., Int. Ed. 1998, 37,
2984-2985.
(37) Frank, W.; Schneider, J.; Muller-Becker, S. Chem. Commun. 1993,
799-800.
(40) Lange’s Handbook of Chemistry, 15th ed.; McGraw-Hill: New York,
1999.
(38) Frank, W.; Muller-Becker, S.; Schneider, J. Z. Anorg. Allg. Chem.
1993, 619, 1073-1082.
(41) Deacon, G. B.; Phillips, R. J. Coord. Chem. ReV. 1980, 33, 227-250.
(42) Soltek, R. Winray GL; Zuru¨k zum Anorganisch-Chemischen Institut:
Heidelberg, 2000.
(39) Frank, W.; Reiland, V. Acta Crystallogr., Sect. C 1998, 54, 1626-
1628.
4204 Inorganic Chemistry, Vol. 41, No. 16, 2002

Molecular Precursors for Ferroelectric Materials
in these studies, mixtures of oxides were obtained for all of
the samples. The higher temperatures required to produce
crystalline samples may favor disproportionation and redis-
tribution of metal atoms in the oxide system and allow the
formation of the stable bismuth-rich perovskite phase in the
case of 3 and 4. In all cases, the conditions required to
produce the bimetallic oxide are substantially milder than
those required under conventional solid-state synthesis
conditions. For example, the formation of a low temperature
form of BiTaO4 with an orthorhombic unit cell (a ) 4.957
Å, b ) 11.763 Å, c ) 5.633 Å) has been reported to occur
previously at 900 °C for 18 h.27 We have obtained the same
material from the conversion of 2a to the oxide through
heating at 700 °C for 3 h.
hydrolysis. The materials can be converted to bimetallic
oxides by heating in an oxygen-containing atmosphere at
600 °C. Powder X-ray diffraction and TGA studies of the
oxides indicate that the ratio of metals in the molecular
precursor influences the ratio of metals in the final oxide.
The titanium containing complexes gave the bismuth-rich
Bi4Ti3O12 phase along with the Bi2Ti4O11 phase expected on
the basis of the metal stoichiometry of the precursor.
Acknowledgment. We would like to thank the Robert
A. Welch foundation and the NSF for support of this work.
J.H.T. would like to thank Dr. Terry Marriott for aid in
collecting the mass spectra of complex 3 and Dr. Larry
Alemany for aid in low temperature NMR experiments and
for helpful discussion.
Conclusion
Supporting Information Available: Full presentation of the
crystallographic data and experimental parameters including atomic
coordinates, bond lengths and angles, anisotropic displacement
parameters, and hydrogen atom coordinates (CIF). Detailed num-
bering schemes for the structures of 1a, 2a, 3, and 4, representative
powder X-ray diffraction patterns, TGA spectra, and IR data. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Bifunctional hydroxy-carboxylate ligands offer a conve-
nient and direct route to the synthesis of heterobimetallic
complexes. We have demonstrated that it is possible to effect
the construction of these compounds in a rational and
stepwise manner by choosing ligands whose functional group
protons have differing acidities. The compounds synthesized
in this manner demonstrate both good solubility in a variety
of organic solvents and excellent stability against unwanted
IC0255454
Inorganic Chemistry, Vol. 41, No. 16, 2002 4205